This proposal is to continue and extend the research begun on my previous grant on different modes of perceptual processing and their dependence on different ways of allocating attention. Attention can be set to select different aspects of the environment, ranging from the global registration of scene properties to the focused analysis of the local conjunctions of features that distinguish individual objects. In the global processing mode, studies of size judgements will include tests for implicit priming of the mean and for illusory representations of the population. Generalization to other dimensions such as color and orientation will be tested, as will various capacity limits inthis global processing mode. Focused attention in feature binding will be tested by varying the structural organization of the display, and testing the use of advance information in search. Some fMRI and TMS studies are also proposed to explore the brain areas involved in different aspects of the binding task, distinguishing spatial scanning, suppression of distractors and feature integration. Tests of a "re-entry" account of binding are also proposed. A third series of experiments explores the maintenance of binding in visual working memory. Experiments are proposed on the retention of object representations for short periods after they disappear, comparing the retention of features and of their bindings and exploring capacity limits and vulnerability to interference. Other studies will explore the relation between long term learning and visual memory to see whether and in what ways the dissociation between short and long term binding that we tentatively identified holds up in other tests. The last project analyses different components of control in switching attention between different attributes of an object - what could be called an "unbinding problem". It will use speeded classification paradigms to identify sources of interference and the components of control and to locate them in the brain through fMRI. Replies to the questions about visual attention and memory should help to determine the architecture of the systems underlying performance in many real world tasks, They may also provide some understanding of how they fail in normal people in conditions of overload, or in patients wtih brain injuries causing neglect or Balint's syndrome. [unreadable] [unreadable]